Part 5 of 6 · The test

Why a single blood test cannot show that anything worked

In a person whose body has not changed at all, repeat hs-CRP results scatter by tens of per cent. The before-and-after test sold across this category measures that scatter.

The before-and-after blood test is sold across this whole category as the proof that a protocol worked. The data on how much that test moves on its own turns the idea inside out.

BModerate evidence

0.44

Median within-person coefficient of variation for hs-CRP, across 60 studies of biological variation (range 0.27–0.76).

Source: Gough 2024, PMID 39485740

In a person whose underlying state has not changed at all, repeat measurements scatter by tens of per cent as a matter of routine. A lower second result can be nothing more than the same person measured twice.

So how large does a change have to be before it carries any information? That threshold can be calculated, and it is larger than most people expect.

Open the chapter this comes from

Behind the paywall · part 5

The threshold, and when testing is still worth it

The reference change value: how large a change must be to mean anything
Why there is no such thing as “an inflammation level”
The two situations where a test does carry information
What would make a self-test informative, and what it costs

$39

USD · one payment · no subscription

All nine chapters and all 26 graded claims
Web reader and a downloadable PDF
Every future correction and revision, included

I accept the Terms of Service and the Refund Policy, and I have read the Right of Withdrawal notice.I request immediate access to the digital content and I acknowledge that I will lose my right of withdrawal once access begins.

Tick both boxes above to enable checkout.

Already purchased? Sign in with the purchase email